How Fast Was the Mitsubishi Zero?

[Venturi]

Excellent article discussing the performance of the remarkable A6M2, and comparing it to some of its contemporary fighter aircraft. Including the F4F and the P-40.

 

http://www.pacificaviationmuseum.org/pearl-harbor-blog/p30/

 

Edited August 25, 2017 by Venturi

[=362nd_FS=Hiromachi]

I usually refer to this article: http://www.j-aircraft.com/research/rdunn/zeroperformance/zero_performance.htm

And this since it has some relatively well documented numbers: http://www.wwiiaircraftperformance.org/japan/a6m2-oct2342.pdf

In general it took me few years of research and I still cannot establish anything beyond above assumptions, even despite considerable effort and spendings on materials. Major issue is absence in any major museum or archive translations (or even non translated) of the flight manuals. I am aware of three documents directly referring to Zero performance - a 1940 12-Shi pilot notes or 1941 flight manual but those were only listed as documents not translated by Allies and either destroyed or returned to Japan. I have only one such document,  JICPOA translation of October 1943 general notes on various Zero models, which lists A6M2 top speed at 275 knots. This corresponds with Sakai reference to speed at maximum continous power. 

I also have U.S. card from Informational Intelligence Summary No. 44-11 giving maximum established speed by them in tests and maximum speed they thought (or calculated) that could be achieved in an aircraft at best condition, corresponding with Sakai note of 345 mph. Sakai's note is also confirmed by another pilot, Saburo Horita, who was shot down in Solomons campaign : https://theamericanwarrior.com/2014/12/11/warrior-adversary-saburo-horitas-story/

(I have his interrogation report).

 

I wanted to run a simulation based on hard data such as drag coefficient, prop. efficiency, engine power ratings and other given numbers but so far found nobody who could help with that.

[Venturi]

I have Goodwin's "Japanese Aero-Engines" already, which I believe is probably the authoritative source to date on Japanese engines including the Sakae 12 radial. The authors are or were extremely active at j-aircraft.com. 

 

Yes, it sounds like the A6M2 could do about 345mph at best altitude with "overboost". However, how long it was safe to "overboost", and what this actually means in terms of manifold pressure?

To my knowledge, we don't have an official top speed with the P-40E doing 56"/3000rpm - which would raise its HP from 1150 to 1475. I do know it was faster than the A6M2, even at 1150hp - 355mph. 

Edited August 25, 2017 by Venturi

[=362nd_FS=Hiromachi]

I have power curves for Sakae 12 along with cooling, supercharger and other graphs. It's more than enough to work with.

 

Overboost would be considered beyond rated power. Based on this: http://www.enginehistory.org/Piston/Japanese/JapaneseEngTesting/JapaneseEngTesting.shtml emergency (or as others call it overboost) was permitted for 3 minutes, but from readings I know pilots disregarded this limit and were overboosting for much longer periods. The way this system worked was relatively simple, there was a knob on front panel which pilot had to pull. That removed the limitations inposted on manifold pressure and pilot could reach take-off/emergency power or perhaps even more.

Take-off/Emergency power would be + 250 mmHg manifold pressure / 2550 rpm, whereas rated power would be +150 mmHg manifold pressure / 2500 rpm.

When I get home I will check this to be sure, but for now I'm stuck at work.

[=362nd_FS=Hiromachi]

Alright, I'm back at home. 

So here are some basic data from manuals and other sources:

Take-off/Emergency - 940 HP / +250 mmHg / 2550 RPM

Rated power - 830 HP / +150 mmHg / 2500 RPM

                       950 HP / +150 mmHg / 2500 RPM at rated altitude (4200 m)

Normal power - 700 HP / +50 mmHg / 2350 RPM

                          820 HP / +50 mmHg / 2350 RPM at 4200 m

 

Propeller - Sumitomo Hamilton standard type, 2.9 m, pitch adjustment - 25 deg - 45 deg, max. prop. efficiency - 0.75 - 0.765

 

Book that I have gives to a Zero (A6M1 in this regard, a prototype with Zuisei engine) a drag coefficient of actual aircraft around 0.018. Japanese version of the same indicates slightly higher number for same model of 0.0200, presumably after corrections (?). For A6M3 given number is 0.0215. 

[JtD]

With 950 hp, 0.75 prop efficiency, 275 knots at 15000feet, 2300kg weight, 0.9 Oswald coefficient, the resulting cdo is 0.020. The numbers add up.

 

Since emergency power is roughly 100 extra hp down low, the engine will maintain this power advantage throughout the altitude, but will not be able to maintain the 250mm boost up to the same 4200m the 150mm are being maintained to. Even with the slightly higher rpm.

 

Assuming that due to increased rpm and higher indicated speed there'd be a stronger ram effect, the new full throttle altitude would be in the region of 2000-3000 feet lower. Interestingly, using the same efficiencies, top speed only increases by a couple of knots. As you go down, air gets thicker, and much of the added power is needed to overcome the increased resistance.

 

The main effect would be that at lower altitudes, you'd get roughly a 10 knot speed increase from using WEP instead of rated power.

 

345 mph would require drag coefficients in the region of 0.016, or much increased prop efficiencies (0.85) in combination with drag coefficients around 0.019. Prop efficiencies of 0.85 are achievable, but I'm not aware of a change in props on the A6M2 at any point.

[Venturi]

Slightly off topic, but I understand props can be tuned for high acceleration, high drag - or low acceleration, low drag. Obviously, prop pitch is the major way this is achieved on variable pitch props. However, I am referring to the prop blade/disc itself. In other words, there is some sort of trade off between maximum speed and maximum acceleration / climb. Can anyone further elucidate this design dynamic?

Edited August 25, 2017 by Venturi

[=362nd_FS=Hiromachi]

Weight would be closer to 2481 kg, at least by the manual, that is for regular condition aircraft. Aspect ratio is 6.42. No idea of Oswald, but in past when trying to calculate performance for A6M3 "we" (me and few other guys working around War Thunder) used number closer to 0.815 and all the other numbers added up. 

For top speed there are two sources I have - 1943 manual gives 275 knots at 4400 meters, 1945 performance chart for all Navy fighters lists it as 275 knots at 5000 meters. 

Plus there is still a matter of this report: http://www.wwiiaircraftperformance.org/japan/a6m2-oct2342.pdf

 

I remember going through Horikoshi book that at the time when they were working on A6M3 they found after carrying flight tests that speed increase from Sakae 21 added power was far smaller than calculated by Mitsubishi and for the first time in company history, theoretical calculations were more optimistic than actual results (in case of A5M or A6M2 recorded in flight results exceeded theoretical calculations) and explanation for this was given to specific air intake design of A6M2. Jiro Horikoshi writes: "In the pre-war days the performance recorded in tests of new fighters usually exceeded the engineers pre-flight estimates. This resulted from the ram effect of the engines air intake which would allow greater output in flight than could be anticipated through restricted medium of bench tests. The A6M3 reversed this trend, and its performance fell below pre flight estimates". A6M3 in this regard had air intake at the top of engine cowling and it was substantially shorter than in A6M2, which is believed to have reduced ram effect in this specific condition.  

 

Anyway, thank you Jtd for trying to hop in and help. I find it difficult to work with A6M2 since there are few well documented numbers, most is based on anecdotal evidence. I will later get results of Army-Navy contest held in 1941 where Zero proved to be superior to both Ki-43 and Ki-44, in fact keeping up with Ki-44 in level flight. But rewriting this stuff will take time.

 

@Venturi I am by no means specialist and if wign airfoils are already a complicated topic for me, propeller blade airfoils are even harder nut to crack. In general blade size, area and thickness will affect the speed and performance. From my readings of IJN documents and Kawanishi - Sumitomo developments around N1K2-J it was found that Navy was producing too thick blades and if it tried to match propeller designs such as ones used in P-51 this could give positive results. A special set of propellers with thinner blades was produced and tested, which allowed N1K2-J to increase top speed by 10-14 knots. 

To get specifics you should ask someone who knows this topic. 

Edited August 25, 2017 by =LD=Hiromachi

[Venturi]

The unfortunate thing is that so many of the Japanese aircraft and engines suffer under this informational limitation. The Mitsubishi Zero and the Sakae 12 being two of the best documented in their respective categories - and as you say, even then there are serious holes in the framework of knowledge.

 

As you mentioned earlier, even the best sources have an incredible amount of speculation regarding the specifics . Mostly resulting from the lack of surviving documentation, the fragmented, rapidly-changing, and craftshop nature of Japanese industry at this time, and of course the lack of surviving examples of many of these engine and aircraft types.

 

As I said, "Japanese Aero-Engines" by Goodwin and j-aircraft.com are probably the two best resources at this time that I know of.

 

This has little bearing on the actual equipment itself. From both contemporary accounts and modern experiences operating the Sakae 12, it was a reliable engine.

Edited August 25, 2017 by Venturi

[=362nd_FS=Hiromachi]

To be fair I have far easier time finding documents for Army aircraft than for Navy. I managed to gather manuals for Ki-43-I, Ki-43-II, Ki-44, Ki-61, Ki-84 and bunch of Army bombers. But until now I did not find one complete flight manual for a Navy aircraft. I could never understand that discrepancy, I know they existed and heck ... I even know titles of some of those, but could not find even one. To be fair though I did not contact any Japanese archive and in particular Mitsubishi musem dedicated to ww2 aircraft developments, but that requires writing in Japanese of which I am not capable of. 

 

Sakae 12 was incredibly reliable. Up to a point when authors of http://www.tainanbooks.com/ were unable to find a single example of aircraft loss due to engine failure for the whole duration of Tainan Ku and Dai-4 Ku operations in New Guinea and Guadalcanal. Engines had rigorous maintenance regimes as well which must have complemented this high reliability. 

[JtD]

In other words, there is some sort of trade off between maximum speed and maximum acceleration / climb.

Not necessarily so, you can have props that are fairly efficient at both low and high speed. In a basic approach, besides thickness, size matters. A small prop (narrow blades, low number of blades, small diameter) is higher loaded than a larger one, which means that the angle of attack of the individual propeller blades has to be larger in comparison (when going at the same speed driven by the same power). The small prop has to deliver more lift (i.e. thrust) per section than the big one, utilizing higher lift coefficients. As with every other airfoil, there is an optimum, most efficient angle of attack, where best lift/drag ratio is achieved. The small prop at low speeds requires inefficiently high angles of attack, while a big prop at high speeds requires inefficiently low angles of attack.

 

With a 2.9m diameter and a .6875(?) reduction ratio the prop of the A6M2 appears to have been extremely lightly loaded, which would explain the poor prop efficiency quoted by Hiromachi.

 

Weight would be closer to 2481 kg, at least by the manual, that is for regular condition aircraft. Aspect ratio is 6.42. No idea of Oswald, but in past when trying to calculate performance for A6M3 "we" (me and few other guys working around War Thunder) used number closer to 0.815 and all the other numbers added up.

Neither Oswald nor weight are overly important for determining cdo at high speed. It would be for estimating climb rates or other, lower speed performances. I just gave it for the complete picture. But if you already know that the figures add up, what are you still looking for?

 

Plus there is still a matter of this report: http://www.wwiiaircraftperformance.org/japan/a6m2-oct2342.pdf

As it says, not corrected for standard atmosphere and not corrected for compressibility. The effect of the first is unknown, the effect of the latter is that speeds are given higher than they really are, by a few mph, except for at sea level. For the top speed stated, about 4-5mph.

[Venturi]

Thanks JtD. I believe there is also the matter of differing "optimal blade AoA" for each forward aircraft speed though - since you mention angle of attack, the "angle" is relatively proportional to the speed at which airflow longitudinal to the aircraft path of flight is approaching the forward aspect of the blade, no?

 

By this reasoning, a prop blade, which is on a plane moving forward at 300mph, will have a lower effective angle of attack for a given prop pitch, than that same blade on a plane moving forward at 180mph, at that identical prop pitch...

 

The Zero's controls became extremely heavy at speeds over 200kts, so I imagine the prop was "loaded lightly" to optimize it for those speeds of flight... if my reasoning is correct. This also accounts for why the prop was relatively inefficient at high speeds.

 

I may be missing something, though?

Edited August 25, 2017 by Venturi

[=362nd_FS=Hiromachi]

Neither Oswald nor weight are overly important for determining cdo at high speed. It would be for estimating climb rates or other, lower speed performances. I just gave it for the complete picture. But if you already know that the figures add up, what are you still looking for?  

 

I knew some of them add up for A6M3 because I had hard numbers for it. For A6M2 its still search up until I find actual manual, for now such calculations help quite a lot.  

 

 

 

The Zero's controls became extremely heavy at speeds over 200kts, so I imagine the prop was "loaded lightly" to optimize it for those speeds of flight... if my reasoning is correct. This also accounts for why the prop was relatively inefficient at high speeds.

Ailerons were heavy. Elevator and rudder were not nearly as close to ailerons, particularly elevator had to go through some changes as well as flexible control cables were used, since in early flights elevator was too light for test pilot preferences, especially at high speeds. 

 

Edit: January 1941 had witnessed the annual competition between the fighter aircraft of the Imperial Army and Navy. The A6M2, piloted alternately by Lieutenant Yoshitomi and Lieutenant Shimokawa, had been entered against the Army Nakajima-built Ki-27, Ki-43-I and Ki-44-I. Following the tests, the Yokosuka Ku provided a summary reports, which were as follows:

1. Speed: Contrary to popular belief prior to the test, the A6M2 fighter proved equal in maximum speed to the Ki-44 fighter, which had a listed speed value 23 m.p.h. greater than that of the Navy plane. The Zeke proved to be considerably faster than the Ki-43 fighter. The top speed recorded by the A6M2 during the competition was 329 mph (530 km/h).
2. Climb: In a steady climb from 42 ft to 13,124 ft the Zeke held a slight advantage over the Oscar, despite the fact that the latter was a lighter aircraft. In zooming performance the Zeke proved far superior to Oscar. The Ki-44 slightly exceeded the Zeke in continuous climbing performance. Both planes were equal in zooming climbs.
3. Turning radius: The Zeke is superior in executing short radius climbing turns to either the Oscar or Ki-44 fighters.
4. Dogfighting: Zeke clearly bested its Army opponents in simulated operational maneuvers. This can be ascribed to the overall aerodynamic refinement of the Zeke fighter.

Following month saw Horikoshi discussing with Lt. Shimokawa and Kofukuda of Yokosuka Ku matters regarding Zekes overall performance capabilities and limitations. Following conclusions were drawn:

1. Superiority of the A6M2 over the Oscar in overall dogfighting characteristics, despite ‘the on-paper superiority’ of the latter one in such respects as the lower weight, lower wing loading, etc. may be ascribed to the Zekes better stability and controllability characteristics, and to its better stall resistance at low speeds. The contest results were unexpected, for on paper Ki-43 possessed a definite advantage over A6M2.
2. The A6M2 clearly bests the Ki-44 in maneuverability. Army fighter’s announced top speed of 339 mph is not correct, the aeroplane could not exceed the Zeke’s maximum speed. This disparity obviously results from difference in performance-test-requirements by the Army and the Navy, the former conducts its test flights under conditions most favorable to maximum performance, while Navy’s aircraft are fully loaded for testing purposes.
3. In level flight Army’s Ki-27 can turn inside A6M2 Zeke but the latter has superior vertical maneuverability. In overall dogfighting characteristics, the Zeke maintains a clear superiority.
4. It is possible to delay undesired spinning and to reduce by 10 % the turning radius of the a6M2, if the pilot makes a slight inward slip during a turn. This slight additional control has proven a definite advantage in combat. 

I'll later get the scan of Intelligence Summary No. 44-11

Edited August 25, 2017 by =LD=Hiromachi

[Holtzauge]

I know I promised to do the A6M2 model in C++ and sorry for not doing this earlier Hiromachi  but better late than never?

 

Anyway, here are the first C++ simulation results: I only have the power/altitude chart you provided so I don’t have any exhaust thrust numbers for the Sakae. OTOH I don’t know how well/if the Zero utilized this effect but I’ve assumed a similar level of contribution like on other planes where I don’t have any specific data. In addition the ram recovery effect on power is another unknown so this is also based on a sort of average of what other planes had.

 

So in order to get the modeling as good as it can get I need known data points for speed and climb in combination with weights. Here it is important to get representative figures, i.e. avoid the outliers since this will otherwise skew the simulations. Once this is in place for a known power then the model can then be used to access other interesting estimates like dive, turn rate, climb time and acceleration performance etc. not only for that specific power but for other weight and power combinations as well.

 

But again, this is just a first shot and I can refine the model if we can decide on which performance estimates and/or flight trial data to base the C++ model on.

Edited August 27, 2017 by Holtzauge

[=362nd_FS=Hiromachi]

I almost forgot about that mate. Thank you for it, even if it's rough and lacks details I havent provided yet. I am currently in process of searching of three documents:

1. Information on operation of type 0 mark I carrier fighter 

2. Information on piloting experimental carrier fighter number 12 

3. Information on piloting type zero carrier fighter 

Those were not translated by JICPOA and their fate is unknown, if I will get my hands on something providing better detail I'll post it here. In regard to weight, as mentioned above by Jiro Horikoshi himself, Navy has conducted all tests in fully loaded condition in standardized manner, unlike Army (or at least Nakajima). So pointed above 2481 kg is a total weight for aircraft fully loaded. 

[Venturi]

Looking at the speed v altitude graph, didn't the Zero have a multi-gear supercharger? 

[Venturi]

Oh never mind - that was the Sakae 21. Guess the critical altitude of the Sakae 12 was about 4.2km, about 13,500ft?

[=362nd_FS=Hiromachi]

Oh never mind - that was the Sakae 21. Guess the critical altitude of the Sakae 12 was about 4.2km, about 13,500ft?

Yes, 4.2 km.